WO2024056064A1 - 转弯路径规划方法、设备、车辆及存储介质 - Google Patents

转弯路径规划方法、设备、车辆及存储介质 Download PDF

Info

Publication number
WO2024056064A1
WO2024056064A1 PCT/CN2023/119059 CN2023119059W WO2024056064A1 WO 2024056064 A1 WO2024056064 A1 WO 2024056064A1 CN 2023119059 W CN2023119059 W CN 2023119059W WO 2024056064 A1 WO2024056064 A1 WO 2024056064A1
Authority
WO
WIPO (PCT)
Prior art keywords
grid
vehicle
turning
path
intersection
Prior art date
Application number
PCT/CN2023/119059
Other languages
English (en)
French (fr)
Inventor
肖智冲
陈建兴
李杨
邱杰
吴浩涛
Original Assignee
广州小鹏自动驾驶科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 广州小鹏自动驾驶科技有限公司 filed Critical 广州小鹏自动驾驶科技有限公司
Publication of WO2024056064A1 publication Critical patent/WO2024056064A1/zh

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
    • G01C21/34—Route searching; Route guidance
    • G01C21/3453—Special cost functions, i.e. other than distance or default speed limit of road segments
    • G01C21/3492—Special cost functions, i.e. other than distance or default speed limit of road segments employing speed data or traffic data, e.g. real-time or historical
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Definitions

  • the present application relates to the field of vehicle technology, and in particular to a turning path planning method, equipment, vehicle and storage medium.
  • intersection scenes In urban autonomous driving scenarios, intersection scenes have extremely complex and particular characteristics, and the smooth traffic capacity of intersection scenes is one of the important evaluation indicators.
  • this application provides a turning path planning method, equipment, vehicle and storage medium, which can generate a planned path through the turning intersection according to the actual traffic flow at the turning intersection, and improve the efficiency of vehicles passing through the turning intersection. security.
  • a first aspect of this application provides a turning path planning method, which method includes:
  • a planned turning path of the ego vehicle is obtained, so that the ego vehicle moves through the target turning intersection according to the planned turning path.
  • the plurality of trajectory points include the starting point of the turning path and the end point of the turning path;
  • Obtaining the planned turning path of the own vehicle based on the plurality of trajectory points of the target vehicle includes:
  • the evaluation function values of the multiple paths between the starting point of the turning path and the end point of the turning path are respectively obtained
  • the path whose evaluation function value meets the preset conditions among the multiple paths is the planned turning path of the own vehicle.
  • the evaluation functions of the multiple paths between the starting point of the turning path and the end point of the turning path are respectively obtained. Values, including:
  • the evaluation function value of each path in the multiple paths is obtained.
  • the reference line is the outer turning boundary line of the target turning intersection, and the outer turning boundary line is obtained by offsetting the outermost lane line of the target turning intersection by a preset distance;
  • the inner turning boundary line of the target turning intersection is a straight line connecting the first boundary point at the innermost entering lane line and the second boundary point at the innermost exiting lane line of the target turning intersection;
  • Constructing the SL grid map within the set range of the target turning intersection includes: constructing the SL grid map between the outer turning boundary line and the inner turning boundary line of the target turning intersection.
  • the set performance index function value of the grid is obtained based on the heading angle offset and/or distance offset between the grid and adjacent grids.
  • determining the weight value of each grid based on the distribution data of the plurality of trajectory points in each grid includes:
  • obtaining the evaluation function value of each path in the multiple paths based on the performance index function value and the weight value of each grid includes:
  • the evaluation function value of each path in the multiple paths is obtained.
  • obtaining the correction coefficient of the weight value of each grid based on the shape data of the target turning intersection and the position data of each grid includes:
  • the offset data of each grid relative to the set safe position is obtained respectively;
  • the correction coefficient of the weight value of each grid is obtained according to the offset data of each grid.
  • obtaining multiple trajectory points of the turning path of the target vehicle in front of the own vehicle in the target turning intersection further includes:
  • the pre-set conditions include some or all of the following: the distance between the vehicle in the same direction and the self-vehicle is not less than the set Fixed distance threshold; the difference between the heading angle of the vehicle in the same direction and the heading angle of the own vehicle is less than the set angle threshold; the speed of the vehicle in the same direction is not less than the set speed threshold.
  • the method further includes: smoothing the planned turning path of the own vehicle through a conjugate gradient algorithm.
  • a second aspect of this application provides a computing device, including:
  • a memory has executable code stored thereon, and when the executable code is executed by the processor, causes the processor to perform the method as described above.
  • a third aspect of the present application provides a vehicle, including the computing device as described above.
  • a fourth aspect of the present application provides a computer-readable storage medium on which executable code is stored.
  • the processor is caused to execute the method as described above.
  • the technical solution of this application obtains the planned turning path of the own vehicle based on the multiple trajectory points of the turning path of the target vehicle in front of the own vehicle passing through the target turning intersection, so that the own vehicle moves through the target turning intersection according to the planned turning path; It is not necessary to identify the lane lines of the target turning intersection.
  • the turning path can be calculated based on the actual traffic flow at the target turning intersection and the multiple trajectory points of the turning path of the target vehicle in front of the vehicle passing through the target turning intersection. Plan to obtain the planned turning path of the own vehicle.
  • the turning path of the target vehicle in the target turning intersection can avoid dangerous areas or obstacles in the target turning intersection. When the own vehicle passes through the target turning intersection according to the planned turning path, it can also avoid Navigating dangerous areas or obstacles in the target turning intersection can avoid dangerous cuts or collisions of your own vehicle and improve the safety of vehicles passing through the target turning intersection.
  • Figure 1 is a schematic flowchart of a turning path planning method according to an embodiment of the present application
  • Figure 2 is a schematic flowchart of a turning path planning method according to another embodiment of the present application.
  • Figure 3 is a schematic diagram of the scene in the embodiment of Figure 2;
  • Figure 4 is a schematic diagram of the SL grid diagram and correction coefficient curve according to an embodiment of the present application.
  • Figure 5 is a schematic structural diagram of a computing device according to an embodiment of the present application.
  • first, second, third, etc. may be used in this application to describe various information, the information should not be limited to these terms. These terms are only used to distinguish information of the same type from each other.
  • first information may also be called second information, and similarly, the second information may also be called first information. Therefore, features defined as “first” and “second” may explicitly or implicitly include one or more of these features.
  • plurality means two or more than two, unless otherwise explicitly and specifically limited.
  • embodiments of the present application provide a turning path planning method, which can generate a planned path through the turning intersection based on the actual traffic flow at the turning intersection, thereby improving the safety of vehicles passing through the turning intersection.
  • Figure 1 is a schematic flowchart of a turning path planning method according to an embodiment of the present application.
  • a turning path planning method includes:
  • the target turning intersection can be a single left-turn intersection, a single right-turn intersection, or a two-way turn intersection that can both turn left and right.
  • the trajectory points of the target vehicle in front of the self-driving vehicle and within the target turning intersection are collected by setting the device, and the trajectory points of the target vehicle in front of the self-driving vehicle at the target turning intersection are obtained. Multiple trajectory points of the turning path within the curved intersection.
  • the planned turning path of the own vehicle is obtained based on multiple trajectory points of the target vehicle, so that the own vehicle moves through the target turning intersection according to the planned turning path.
  • the multiple trajectory points of the target vehicle include the starting point and the end point of the turning path; obtaining the planned turning path of the own vehicle based on the multiple trajectory points of the target vehicle includes: determining the reference line of the Frenet coordinate system, where, The reference line is the turning boundary line within the target turning intersection; according to the reference line, multiple trajectory points are converted from the current coordinate system to the Frenet coordinate system; in the Frenet coordinate system, according to multiple trajectory points and the preset evaluation function, respectively Obtain the respective evaluation function values of multiple paths between the starting point of the turning path and the end point of the turning path; based on the respective evaluation function values of the multiple paths, determine that the path whose evaluation function value meets the preset conditions among the multiple paths is the planned turn of the own vehicle path so that the vehicle can move through the target turning intersection according to the planned turning path.
  • the starting point and end point of the planned turning path can be obtained based on multiple trajectory points of the target vehicle; the turning boundary line on one side of the target turning intersection is determined as the reference line of the Frenet coordinate system, and the multiple trajectory points are Convert from the current coordinate system to the Frenet coordinate system; in the Frenet coordinate system, according to multiple trajectory points, multiple paths between the starting point and the end point of the planned turning path are obtained, and according to the preset evaluation function, multiple paths are obtained respectively Respective evaluation function values; based on the respective evaluation function values of multiple paths, determine the path with the smallest evaluation function value among the multiple paths as the planned turning path of the own vehicle, so that the own vehicle moves through the target turning intersection according to the planned turning path.
  • the planned turning path of the own vehicle is obtained based on multiple trajectory points of the turning path of the target vehicle in front of the own vehicle passing through the target turning intersection, so that the own vehicle moves through the target turning intersection according to the planned turning path; It is not necessary to identify the lane lines of the target turning intersection, but rely on the actual traffic flow of the target turning intersection, and perform turning path planning based on the multiple trajectory points of the target vehicle in front of the vehicle passing through the target turning intersection to obtain the planned turn of the own vehicle.
  • the turning path of the target vehicle in the target turning intersection can avoid dangerous areas or obstacles in the target turning intersection. When the vehicle passes through the target turning intersection according to the planned turning path, it can also avoid dangerous areas in the target turning intersection. or obstacles, which can avoid dangerous cuts or collisions of the own vehicle and improve the safety of the vehicle passing through the target turning intersection.
  • FIG. 2 is a schematic flowchart of a turning path planning method according to another embodiment of the present application.
  • FIG. 3 is a schematic diagram of the scene in the embodiment of FIG. 2 .
  • turning path planning at a left-turn intersection as the target turning intersection is used as an example for explanation.
  • the figure takes two lanes as an example.
  • the intersection frame 301 shows the intersection range of a left-turn intersection.
  • the left-turn intersection has two entry lanes and two exit lanes, namely the first entry lane 321 and the second exit lane. Entering lane 322, first exiting lane 331, and second exiting lane 332.
  • the first boundary line 3211 and the second boundary line 3212 are lane boundary lines of the first entry lane 321 .
  • the second boundary line 3212 and the third boundary line 3213 are lane boundary lines of the second entry lane 322 .
  • the fourth boundary line 3311 and the fifth boundary line 3312 are lane boundary lines of the first exit lane 331 .
  • the fifth boundary line 3312 and the sixth boundary line 3313 are lane boundary lines of the second exit lane 332 .
  • the left-turn intersection also has a first virtual lane line 3041 in the intersection connecting the first entry lane line 3021 and the first exit lane line 3031, and an intersection connecting the second entry lane line 3022 and the second exit lane line 3032.
  • the second virtual lane line 3042 within.
  • the starting point of the first virtual lane line 3041 in the intersection is the intersection P1 of the first entry lane line 3021 and the intersection stop line 305, and the end point is the intersection P2 of the first exit lane line 3031 and the intersection frame 301.
  • the second virtual lane in the intersection The starting point of the line 3042 is the intersection P3 of the second entry lane line 3022 and the intersection stop line 305, and the end point is the intersection P4 of the second exit lane line 3032 and the intersection frame 301.
  • the first virtual lane line 3041 and the second virtual lane line 3042 within the intersection may be generated in advance or generated in real time. According to the embodiment of the present application, the planned turning path through the left-turn intersection can be obtained. It can be understood that the lane line may be but is not limited to the lane center line.
  • a turning path planning method includes:
  • the positioning of the own vehicle can be obtained; according to the positioning of the own vehicle, the specific location of the current road where the own vehicle is located is determined through a preset map, such as a high-precision map.
  • the current road where the vehicle is located includes the target turning intersection (when the vehicle is within a certain range from the target turning intersection), the road information of the target turning intersection can be obtained.
  • the target turning intersection can be a three-way intersection, a crossroads or other types of intersections.
  • the road information of the target turning intersection may include the entry lane line, the exit lane line, the virtual lane line of the intersection, the stop line, the intersection range, the lane boundary line of the entry lane line, the lane boundary line of the exit lane line, etc.
  • the autonomous vehicle can also obtain road information of the target turning intersection through sensing devices (such as cameras and/or radars).
  • the setting range of the target turning intersection can be obtained based on the road information of the target turning intersection, that is, the same-direction vehicle screening range of the target turning intersection can be obtained.
  • the outermost second virtual lane line 3042 in the target turning intersection can be offset by a preset distance to obtain the right boundary line 3061 of the same-direction vehicle screening range to connect the inner boundary line of the first entry lane 321 (
  • the straight line between the first boundary line 3211) and the inner boundary line (fourth boundary line 3311) of the first exit lane 331 is the left boundary line 3062 of the same-direction vehicle screening range.
  • the same-direction vehicles in front of the own vehicle that meet the preset conditions within the same-direction vehicle screening range are determined as target vehicles, and multiple trajectory points of the target vehicles are obtained.
  • the self-vehicle can collect the trajectory information of the same-direction vehicle in the same-direction vehicle screening range through the sensing device.
  • the trajectory information of the same-direction vehicle includes the speed V agent and heading angle ⁇ of the same-direction vehicle. agent , the distance to the own vehicle, trajectory points, etc.; the heading angle ⁇ ego of the own vehicle is obtained through the driving path of the own vehicle.
  • the speed V agent of the same-direction vehicle in the same-direction vehicle screening range the heading angle ⁇ agent , the distance from the own vehicle, and the heading angle ⁇ ego of the own vehicle, the same-direction vehicle in the same direction vehicle screening range is in front of the own vehicle.
  • the vehicles are screened, and the distance between the vehicle in the same direction and the own vehicle is not less than the set distance threshold, the difference between the heading angle ⁇ agent of the vehicle in the same direction and the heading angle ⁇ ego of the own vehicle is less than the set angle threshold, and the speed of the vehicle in the same direction is not less than the set distance threshold.
  • Vehicles in the same direction that are less than the set speed threshold are determined as target vehicles, and multiple trajectory points of the target vehicle are obtained.
  • the same-direction vehicles in front of the own vehicle are screened, and the distance between the center of the same-direction vehicle and the center of the own vehicle is greater than or equal to 3 meters,
  • ⁇ Vehicles in the same direction with ⁇ /2 and speed V agent greater than or equal to 2.0 meters per second are determined as target vehicles, and multiple trajectory points of the target vehicle are obtained.
  • a Frenet coordinate system is established based on the turning boundary line of the target turning intersection.
  • the turning boundary line of the target turning intersection includes an outer turning boundary line and an inner turning boundary line.
  • the outer turning boundary line is obtained by offsetting the outermost lane line of the target turning intersection by a preset distance;
  • the inner turning boundary line of the target turning intersection is The boundary line is a straight line connecting the first boundary point at the innermost entry lane line of the target turning intersection and the second boundary point at the innermost exit lane line.
  • the right boundary line 3061 of the same-direction vehicle screening range can be used as the outer turning boundary line of the target turning intersection
  • the left boundary line 3062 of the same-direction vehicle screening range can be used as the inner turning boundary line of the target turning intersection.
  • the outer turning boundary line of the target turning intersection is used as the reference line of the Frenet coordinate system to establish the Frenet coordinate system.
  • the direction along the reference line is the S axis
  • the direction perpendicular to the reference line is the L axis.
  • a set sampling step is used to sample the same-direction vehicle screening range between the outer turning boundary line and the inner turning boundary line of the target turning intersection, and the outer turning boundary line of the target turning intersection is constructed. and the inside turn boundary line.
  • setting the sampling step size includes setting the sampling step size s along the S-axis direction and setting the sampling step size l along the L-axis direction.
  • the S-axis is used to set the sampling step size s
  • the L-axis is used to set the sampling step size l to screen vehicles in the same direction at the target turning intersection.
  • Select a range for sampling and construct an SL grid map within the set range of the target turning intersection.
  • the SL grid map includes multiple grids, and the size of each grid is s ⁇ l.
  • the sampling step size s is larger than the sampling step size l.
  • the range of the S-axis setting sampling step s can be in the range of 1.0 meters to 5.0 meters, for example, the value is 3.0 meters.
  • the range of the L-axis setting sampling step length l can be in the range of 0.8 meters to 2.0 meters, for example, the value is 0.8 meters.
  • the multiple trajectory points of the target vehicle include the starting point of the turning path and the end point of the turning path.
  • the starting point and end point of the planned turning path can be obtained based on multiple trajectory points of the target vehicle.
  • point P1 or point P3 can be used as the starting point and point P2 or point P4 can be used as the end point according to the position of the vehicle entering the target turning intersection. For example, when the vehicle enters the target turning intersection in the first entry lane 321, point P1 is the starting point of the planned turning path, and point P2 is the end point of the planned turning path.
  • the Dijkstra algorithm can be used to search in the SL grid diagram along the S-axis direction of the Frenet coordinate system to obtain multiple search paths between the starting point of the turning path and the end point of the turning path; For example, the center point of the grid passed by the search path can be used as the trajectory point of the search path.
  • the set performance index function value of the grid is obtained based on the heading angle offset and/or distance offset between the grid and adjacent grids.
  • the set performance index function value of the grid is also called the cost of the grid (cost), including the heading cost (heading cost) and the distance cost (distance cost) of the grid.
  • the cost of the grid is represented by G gird means that the heading angle cost is represented by G hc and the distance cost is represented by G dc .
  • the cost of the grid G gird can be obtained at least based on the heading angle cost G hc and the distance cost G dc .
  • the heading angle cost of each grid on each path in multiple paths can be obtained based on the heading angle offset between this grid and the next adjacent grid; the distance cost of each grid on each path in multiple paths, It can be obtained based on the distance offset between this grid and the next adjacent grid.
  • the SL coordinates of the two grid center points are converted into XY coordinates of the Cartesian coordinate system, and the vector angle formed by the two center points is used as the heading angle offset between the two grids.
  • the Euclidean distance between two grid center points is used as the distance offset between the two grids.
  • the weight value of each grid is determined based on the distribution data of multiple trajectory points in each grid.
  • the weight value of each grid in the search path can be set according to the distribution data of multiple trajectory points in each grid. For example, the weight of a grid with trajectory points is set to be higher than that of a grid without trajectory points. The weight; for another example, set the weight of a grid with more trajectory points to be higher than the weight of a grid with few or no trajectory points.
  • the grid cost is obtained based on the heading angle cost G hc and the distance cost G dc , and the path with the smallest weighted sum of costs of each grid among multiple paths can be determined as the planned turning path of the own vehicle.
  • the weight value of the grid with trajectory points is set to be smaller than the weight value of the grid without trajectory points; and /Or, set the weight value of a grid with many trajectory points to be smaller than the weight value of a grid with few or no trajectory points.
  • multiple trajectory points of the target vehicle are converted from the current coordinate system to the Frenet coordinate system according to the reference line of the Frenet coordinate system.
  • the converted multiple trajectory points are distributed in the grid of the SL grid diagram.
  • the distribution data of multiple trajectory points within each grid can be obtained.
  • Figure 4 shows an example of the SL grid diagram.
  • the dot 401 represents the trajectory point of the target vehicle.
  • Some grids in the SL grid diagram do not have the trajectory points of the target vehicle, and some do.
  • the weight value of the grid can also be called the attenuation factor F agent of the grid.
  • the attenuation factor F agent is a coefficient less than 1 or equal to 1. .
  • the attenuation factor F agent of the grid with the trajectory points of the target vehicle is less than 1, and the attenuation of the grid without the trajectory points of the target vehicle
  • the factor F agent is 1.0.
  • the attenuation factor F agent of a grid with at least two trajectory points of the target vehicle is 0.6
  • the attenuation factor F agent of a grid with one trajectory point of the target vehicle is 0.8
  • the attenuation factor F agent of a grid without a trajectory point of the target vehicle is 0.8.
  • the grid of points has an attenuation factor F agent of 1.0.
  • G gird-i F agent-i *(G hci +G dci ), where F agent-i , G hci , and G dci are respectively the attenuation factor, heading angle cost, and distance cost of the i-th grid in the SL grid map.
  • the correction coefficient of the weight value of each grid is obtained based on the shape data of the target turning intersection and the position data of each grid.
  • the proportional relationship between the length K and width W of the target turning intersection and the SL grid diagram can be used.
  • the L-axis data of the grid is used to obtain the offset distance of each grid relative to the set safe position; the correction coefficient of the weight value of each grid is obtained based on the offset distance of each grid.
  • a correction coefficient curve 402 is designed along the L-axis 403 direction of the Frenet coordinate system.
  • the correction coefficient can give a higher weight to the grid closer to the set safe position.
  • the correction coefficient of the i-th grid of the SL grid diagram is F Li .
  • the corrected weighted cost G gird-i F Li *F agent- i *(G hci +G dci ).
  • the correction coefficient F Li may be a coefficient greater than or equal to 1, and the closer the grid is to the set safety center, The correction coefficient is closer to 1.
  • the correction coefficient curve 402 is a near-parabola along the L-axis direction of the Frenet coordinate system, and its center corresponds to the L-axis center position of the SL grid diagram, which can be expressed as Lx/2, and the extreme value corresponding to the center position is 1.0. Since the shapes of target turning intersections on different roads are different, the length-to-width ratio of the intersection can be used to adjust the center position of the correction coefficient curve.
  • the correction coefficient F Li of the weight value of the i-th grid of the SL grid diagram can be calculated by the following formula:
  • Li is the L-axis data of the i-th grid of the SL grid diagram, which can be obtained by multiplying i by the L-axis sampling step l; is the offset distance of the i-th grid relative to the set safe position; Lx/2 is the center position of the L axis of the SL grid diagram.
  • each of the multiple paths passes through multiple grids in the SL grid diagram, connecting the starting point and the end point of the planned turning path.
  • the set performance index function values of each grid on each path are weighted and added, and the weighted sum of the set performance index function values of each grid on each path is the evaluation function value of each path in the multiple paths.
  • the weighted cost sum ⁇ G gird of the N grids is the path value. Evaluation function value.
  • the path with the smallest evaluation function value can be selected as the planned turning path of the Frenet coordinate system where the vehicle passes through the target turning intersection, so that the vehicle can make the desired turn according to the plan.
  • the turn path moves through the target turn intersection.
  • the planned turning path in the Frenet coordinate system can be converted into the planned turning path in the Cartesian coordinate system to obtain the planned turning path in the Cartesian coordinate system, so that the vehicle can move through the target turning intersection according to the planned turning path in the Cartesian coordinate system.
  • a smoothing algorithm can be used to smooth the planned turning path of the own vehicle to obtain a smooth planned turning path through the target turning intersection.
  • the planned turning path of the own vehicle can be smoothed by including but not limited to the CG (Conjugate Gradient, conjugate gradient) algorithm to obtain a smooth planned turning path, so that the own vehicle can move according to the smooth planned turning path. Move smoothly through targeted turns.
  • CG Conjugate Gradient, conjugate gradient
  • the planned turning path of the own vehicle is obtained based on multiple trajectory points of the turning path of the target vehicle in front of the own vehicle passing through the target turning intersection, so that the own vehicle moves through the target turning intersection according to the planned turning path; It is not necessary to identify the lane lines of the target turning intersection.
  • the turning path planning is performed based on the multiple trajectory points of the turning path of the target vehicle in front of the vehicle passing through the target turning intersection, and the planned turning path of the own vehicle is obtained.
  • the turning path of the target vehicle in the target turning intersection can avoid dangerous areas or obstacles in the target turning intersection. When the vehicle passes through the target turning intersection according to the planned turning path, it can also avoid dangerous areas or obstacles in the target turning intersection. It can avoid dangerous cuts or collisions of your own vehicle and improve the safety of vehicles passing through target turning intersections.
  • the turning boundary line within the target turning intersection is used as the reference line of the Frenet coordinate system to construct the Frenet coordinate system;
  • the SL grid diagram within the set range of the target turning intersection is constructed in the Frenet coordinate system, Convert the multiple trajectory points of the target vehicle from the current coordinate system to the Frenet coordinate system.
  • the multiple trajectory points are distributed in the grid of the SL grid diagram.
  • the distribution data of the multiple trajectory points in each grid of the SL grid diagram Determine the weight value of each grid, and set the weight value of the grid with trajectory points to be smaller than the weight value of the grid without trajectory points; and/or, set the weight value of the grid with many trajectory points to be smaller than the weight value of the grid with few or no trajectory points.
  • the weight value of the grid with the trajectory points through the setting of the weight value of each grid, reduces the cost of each grid with the trajectory points of the target vehicle, so that the planned turning path of the own vehicle passes through the trajectory points of the target vehicle as much as possible. It can avoid dangerous cuts or collisions of your own vehicle and improve the safety of vehicles passing through target turning intersections.
  • the offset data of each grid relative to the set safe position is obtained; according to each grid
  • the correction coefficient of each grid is obtained from the offset data.
  • the correction coefficient is a coefficient greater than 1. The closer the correction coefficient of each grid is to the center of the target turning intersection, the smaller the correction coefficient is, so that the cost of each grid closer to the center is smaller.
  • the closer the planned turning path of the own vehicle is to the center of the set range of the target turning intersection it can avoid that the planned turning path of the own vehicle is located at the boundary of the target turning intersection.
  • the own vehicle moves through the target turning intersection according to the planned turning path, it can Deviating from the boundaries on both sides of the target turning intersection can avoid dangerous cuts or collisions of the own vehicle and improve the safety of the vehicle passing through the target turning intersection.
  • this application also provides a computing device and corresponding embodiments.
  • FIG. 5 is a schematic structural diagram of a computing device according to an embodiment of the present application.
  • the computing device may be one or more computer terminals or a server, or a combination of computer terminals and servers, or the like. It can be understood that the server can be a physical server or a logical server virtualized by multiple physical servers. The server can also be a server group composed of multiple servers that can communicate with each other, and each functional module can be distributed on each server in the server group.
  • the computing device is a vehicle-mounted electronic device, which may be, for example, but not limited to, a vehicle's electronic control unit, an autonomous driving system controller, an intelligent navigation device, a smartphone, a smart tablet device and other mobile devices.
  • computing device 500 includes memory 501 and processor 502 .
  • the computing device 500 includes a processor 502 and a memory 501 storing computer programs.
  • the processor 502 executes the stored computer program, the turning path planning method can be implemented.
  • the computing device 500 may be one or more computer terminals, a server, or a combination of computer terminals and servers, or the like.
  • the server can be a physical server or a logical server virtualized by multiple physical servers.
  • the server can also be a server group composed of multiple servers that can communicate with each other, and each functional module can be distributed on each server in the server group.
  • the computing device 500 is a vehicle-mounted electronic device, which may be, for example, but not limited to, a vehicle's electronic control unit, an autonomous driving system controller, a smart navigation device, a smart phone, a smart tablet device and other mobile devices.
  • the processor 502 can be a central processing unit (Central Processing Unit, CPU), or other general-purpose processor, a digital signal processor (Digital Signal Processor, DSP), an application specific integrated circuit (Application Specific Integrated Circuit, ASIC), or an on-site processor.
  • Programmable gate array Field-Programmable Gate Array, FPGA or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.
  • a general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc.
  • Memory 501 may include various types of storage units, such as system memory, read-only memory (ROM), and persistent storage. Among them, ROM can store static data or instructions required by the processor 502 or other modules of the computer. Persistent storage may be readable and writable storage. Persistent storage may be a non-volatile storage device that does not lose stored instructions and data even when the computer is powered off. In some embodiments, the permanent storage device uses a large-capacity storage device (eg, magnetic or optical disk, flash memory) as the permanent storage device. In other embodiments, the permanent storage device may be a removable storage device (eg, floppy disk, optical drive).
  • System memory can be a read-write storage device or a volatile read-write storage device, such as dynamic random access memory.
  • System memory can store some or all of the instructions and data the processor needs to run.
  • memory 501 may include any combination of computer-readable storage media, including various types of semiconductor memory chips (eg, DRAM, SRAM, SDRAM, flash memory, programmable read-only memory), and magnetic and/or optical disks may also be used.
  • memory 501 may include a readable and/or writable removable storage device, such as a compact disc (CD), a read-only digital versatile disc (eg, DVD-ROM, dual-layer DVD-ROM), Read-only Blu-ray discs, ultra-density discs, flash memory cards (such as SD cards, min SD cards, Micro-SD cards, etc.), magnetic floppy disks, etc.
  • a readable and/or writable removable storage device such as a compact disc (CD), a read-only digital versatile disc (eg, DVD-ROM, dual-layer DVD-ROM), Read-only Blu-ray discs, ultra-density discs, flash memory cards (such as SD cards, min SD cards, Micro-SD cards, etc.), magnetic floppy disks, etc.
  • Computer-readable storage media do not contain carrier waves and transient electronic signals that are transmitted wirelessly or wired.
  • the memory 501 stores executable code.
  • the processor 502 can be caused to execute part or all of the above-mentioned methods.
  • the present application also provides a vehicle including the computing device 500 as described above.
  • the method according to the present application can also be implemented as a computer program or computer program product, which computer program or computer program product includes computer program code instructions for executing part or all of the steps in the above method of the present application.
  • the application may also be implemented as a computer-readable storage medium (or a non-transitory machine-readable storage medium or a machine-readable storage medium) with executable code (or computer program or computer instruction code) stored thereon,
  • executable code or computer program or computer instruction code
  • the processor of the vehicle or computing device, server, etc.
  • the processor is caused to perform part or all of the various steps of the above-described method according to the present application.

Landscapes

  • Engineering & Computer Science (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Automation & Control Theory (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Traffic Control Systems (AREA)

Abstract

一种转弯路径规划方法、设备、车辆及存储介质。方法包括：获得自车前方的目标车辆在目标转弯路口内的转弯路径的多个轨迹点（S110）；根据目标车辆的多个轨迹点，获得自车的规划转弯路径，以使自车根据规划转弯路径移动通过目标转弯路口（S120）。

Description

转弯路径规划方法、设备、车辆及存储介质
相关申请
本申请要求于2022年9月16日申请的、申请号为202211125386.X的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及车辆技术领域，尤其涉及一种转弯路径规划方法、设备、车辆及存储介质。
背景技术
在城市自动驾驶场景中，路口场景有着极其复杂的特殊性，路口场景通畅的通行能力是重要的评估指标之一。
相关技术基于高精度地图的自动驾驶，由于实际车流与路口内的车道线存在不符的情况，自动驾驶车辆沿着高精度地图的车道线行驶，会与路口内的实际车流的行为不符，导致危险切入或者碰撞。
发明内容
为解决或部分解决相关技术中存在的问题，本申请提供一种转弯路径规划方法、设备、车辆及存储介质，能够根据转弯路口的实际车流，生成通过转弯路口的规划路径，提高车辆通过转弯路口的安全性。
本申请第一方面提供一种转弯路径规划方法，所述方法包括：
获得自车前方的目标车辆在目标转弯路口内的转弯路径的多个轨迹点；
根据所述目标车辆的所述多个轨迹点，获得所述自车的规划转弯路径，以使所述自车根据所述规划转弯路径移动通过所述目标转弯路口。
一实施例中，所述多个轨迹点包括转弯路径起点和转弯路径终点；
所述根据所述目标车辆的所述多个轨迹点，获得所述自车的规划转弯路径，包括：
确定Frenet坐标系的参考线，其中，所述参考线为所述目标转弯路口内的转弯边界线；
根据所述参考线，将所述多个轨迹点从当前坐标系转换到Frenet坐标系；
在所述Frenet坐标系下，根据所述多个轨迹点和预设的评价函数，分别获得所述转弯路径起点和转弯路径终点之间的多条路径各自的评价函数值；
根据所述多条路径各自的评价函数值，确定所述多条路径中评价函数值符合预设条件的路径为所述自车的规划转弯路径。
一实施例中，所述在所述Frenet坐标系下，根据所述多个轨迹点和预设的评价函数，分别获得所述转弯路径起点和转弯路径终点之间的多条路径各自的评价函数值，包括：
构建所述目标转弯路口的设定范围内的SL网格图，所述多个轨迹点分布于所述SL网格图的网格内；
根据预设路径搜索算法，在所述SL网格图中获得所述转弯路径起点和转弯路径终点之间的多条路径；
获得所述多条路径中的每条路径的评价函数值，包括：
分别获得所述路径上各网格的设定性能指标函数值；
根据所述多个轨迹点在所述各网格内的分布数据分别确定所述各网格的权重值；
根据所述各网格的所述性能指标函数值和所述权重值，获得所述多条路径中的每条路径的评价函数值。
一实施例中，所述参考线为所述目标转弯路口的外侧转弯边界线，所述外侧转弯边界线为将所述目标转弯路口的最外侧车道线偏移预设距离获得；
所述目标转弯路口的内侧转弯边界线为连接所述目标转弯路口的最内侧驶入车道线处的第一边界点和最内侧驶出车道线处的第二边界点的直线；
所述构建所述目标转弯路口的设定范围内的SL网格图，包括：构建所述目标转弯路口的外侧转弯边界线和内侧转弯边界线之间的SL网格图。
一实施例中，所述网格的设定性能指标函数值是根据所述网格与相邻网格之间的航向角偏移和/或距离偏移获得的。
一实施例中，所述根据所述多个轨迹点在所述各网格内的分布数据确定所述各网格的权重值，包括：
设置具有轨迹点的网格的权重高于不具有轨迹点的网格的权重；和/或，
设置轨迹点多的网格的权重高于轨迹点少或不具有轨迹点的网格的权重。
一实施例中，所述根据所述各网格的所述性能指标函数值和所述权重值，获得所述多条路径中的每条路径的评价函数值，包括：
根据所述目标转弯路口的形状数据及各网格的位置数据，获得所述各网格的权重值的修正系数；
根据所述各网格的所述性能指标函数值、所述权重值和所述修正系数，获得所述多条路径中的每条路径的评价函数值。
一实施例中，所述根据所述目标转弯路口的形状数据及所述各网格的位置数据，获得所述各网格的权重值的修正系数，包括：
根据所述目标转弯路口的长度和宽度的比例关系、及所述各网格的L轴数据，分别获得所述各网格相对于设定安全位置的偏移数据；
根据所述各网格的所述偏移数据获得所述各网格的权重值的修正系数。
一实施例中，所述获得自车前方的目标车辆在目标转弯路口内的转弯路径的多个轨迹点，还包括：
将自车前方处于设定路口范围内、且符合预设条件的同向车辆确定为目标车辆，所述符合预设条件包括以下部分或全部：同向车辆与自车之间的距离不小于设定距离阈值；同向车辆的航向角与自车的航向角之差小于设定角度阈值；同向车辆的速度不小于设定速度阈值。
一实施例中，所述方法还包括：通过共轭梯度算法对所述自车的规划转弯路径进行平滑处理。
本申请第二方面提供一种计算设备，包括：
处理器；以及
存储器，其上存储有可执行代码，当所述可执行代码被所述处理器执行时，使所述处理器执行如上所述的方法。
本申请第三方面提供一种车辆，包括如上所述的计算设备。
本申请第四方面提供一种计算机可读存储介质，其上存储有可执行代码，当所述可执行代码被处理器执行时，使所述处理器执行如上所述的方法。
本申请提供的技术方案可以包括以下有益效果：
本申请的技术方案，根据自车前方的目标车辆通过目标转弯路口的转弯路径的多个轨迹点，获得自车的规划转弯路径，以使自车根据规划转弯路径移动通过目标转弯路口；自车可以不必识别目标转弯路口的车道线，根据目标转弯路口的实际车流，根据自车前方的目标车辆通过目标转弯路口的转弯路径的多个轨迹点进行转弯路径规 划，获得自车的规划转弯路径，目标转弯路口内目标车辆行驶的转弯路径可以避开目标转弯路口中的危险地段或障碍物，自车在根据规划转弯路径通过目标转弯路口时，也能避开目标转弯路口中的危险地段或障碍物，能够避免自车的危险切入或者碰撞，提高车辆通过目标转弯路口的安全性。
应当理解的是，以上的一般描述和后文的细节描述仅是示例性和解释性的，并不能限制本申请。
附图说明
通过结合附图对本申请示例性实施方式进行更详细地描述，本申请的上述以及其它目的、特征和优势将变得更加明显，其中，在本申请示例性实施方式中，相同的参考标号通常代表相同部件。
图1是本申请一实施例的转弯路径规划方法的流程示意图；
图2是本申请另一实施例的转弯路径规划方法的流程示意图；
图3是图2实施例的场景示意图；
图4是本申请一实施例的SL网格图和修正系数曲线的示意图；
图5是本申请一实施例的计算设备的结构示意图。
具体实施方式
下面将参照附图更详细地描述本申请的实施方式。虽然附图中显示了本申请的实施方式，然而应该理解，可以以各种形式实现本申请而不应被这里阐述的实施方式所限制。相反，提供这些实施方式是为了使本申请更加透彻和完整，并且能够将本申请的范围完整地传达给本领域的技术人员。
在本申请使用的术语是仅仅出于描述特定实施例的目的，而非旨在限制本申请。在本申请和所附权利要求书中所使用的单数形式的“一种”、“所述”和“该”也旨在包括多数形式，除非上下文清楚地表示其他含义。还应当理解，本文中使用的术语“和/或”是指并包含一个或多个相关联的列出项目的任何或所有可能组合。
应当理解，尽管在本申请可能采用术语“第一”、“第二”、“第三”等来描述各种信息，但这些信息不应限于这些术语。这些术语仅用来将同一类型的信息彼此区分开。例如，在不脱离本申请范围的情况下，第一信息也可以被称为第二信息，类似地，第二信息也可以被称为第一信息。由此，限定有“第一”、“第二”的特征可以明示或者隐含地包括一个或者更多个该特征。在本申请的描述中，“多个”的含义是两个或两个以上，除非另有明确具体的限定。
相关技术基于高精度地图的自动驾驶，由于实际车流与路口内的车道线存在不符的情况，自动驾驶车辆沿着高精度地图的车道线行驶，会与路口内的实际车流的行为不符，导致危险切入或者碰撞，车辆通过路口时存在安全隐患。
针对上述问题，本申请实施例提供一种转弯路径规划方法，能够根据转弯路口的实际车流，生成通过转弯路口的规划路径，提高车辆通过转弯路口的安全性。
以下结合附图详细描述本申请实施例的技术方案。
图1是本申请一实施例的转弯路径规划方法的流程示意图。
参见图1，一种转弯路径规划方法，包括：
在S110中，获得自车前方的目标车辆在目标转弯路口内的转弯路径的多个轨迹点。
可以理解的，目标转弯路口可以是单左转弯路口、单右转弯路口、或者也可以是既可左转又可右转的可双向转弯路口。
一实施例中，在自动驾驶的自车通过目标转弯路口时，通过设定设备对自车前方的、目标转弯路口内目标车辆的轨迹点进行采集，获得自车前方的目标车辆在目标转 弯路口内的转弯路径的多个轨迹点。
在S120中，根据目标车辆的多个轨迹点，获得自车的规划转弯路径，以使自车根据规划转弯路径移动通过目标转弯路口。
一实施例中，目标车辆的多个轨迹点包括转弯路径起点和转弯路径终点；根据目标车辆的多个轨迹点，获得自车的规划转弯路径，包括：确定Frenet坐标系的参考线，其中，参考线为目标转弯路口内的转弯边界线；根据参考线，将多个轨迹点从当前坐标系转换到Frenet坐标系；在Frenet坐标系下，根据多个轨迹点和预设的评价函数，分别获得转弯路径起点和转弯路径终点之间的多条路径各自的评价函数值；根据多条路径各自的评价函数值，确定多条路径中评价函数值符合预设条件的路径为自车的规划转弯路径，以使自车根据规划转弯路径移动通过目标转弯路口。
一实施例中，可以根据目标车辆的多个轨迹点，获得规划转弯路径的起点和终点；将目标转弯路口内其中一侧的转弯边界线确定为Frenet坐标系的参考线，将多个轨迹点从当前坐标系转换到Frenet坐标系；在Frenet坐标系下，根据多个轨迹点，分别获得规划转弯路径的起点和终点之间的多条路径，根据预设的评价函数，分别获得多条路径各自的评价函数值；根据多条路径各自的评价函数值，确定多条路径中评价函数值最小的路径为自车的规划转弯路径，以使自车根据规划转弯路径移动通过目标转弯路口。
本申请实施例中，根据自车前方的目标车辆通过目标转弯路口的转弯路径的多个轨迹点，获得自车的规划转弯路径，以使自车根据规划转弯路径移动通过目标转弯路口；自车可以不必识别目标转弯路口的车道线，而是依靠目标转弯路口的实际车流，根据自车前方的目标车辆通过目标转弯路口的转弯路径的多个轨迹点进行转弯路径规划，获得自车的规划转弯路径，目标转弯路口内目标车辆行驶的转弯路径可以避开目标转弯路口中的危险地段或障碍物，自车在根据规划转弯路径通过目标转弯路口时，也能避开目标转弯路口中的危险地段或障碍物，能够避免自车的危险切入或者碰撞，提高车辆通过目标转弯路口的安全性。
图2是本申请另一实施例的转弯路径规划方法的流程示意图。图3是图2实施例的场景示意图。
结合图3，本实施例中，以在左转弯路口为目标转弯路口进行转弯路径规划为例进行说明。图中以二车道为例，其中，路口框301示出左转弯路口的路口范围，该左转弯路口具有两条驶入车道和两条驶出车道，分别为第一驶入车道321、第二驶入车道322、第一驶出车道331、第二驶出车道332。第一边界线3211和第二边界线3212为第一驶入车道321的车道边界线。第二边界线3212和第三边界线3213为第二驶入车道322的车道边界线。第四边界线3311和第五边界线3312为第一驶出车道331的车道边界线。第五边界线3312和第六边界线3313为第二驶出车道332的车道边界线。该左转弯路口还具有连接第一驶入车道线3021和第一驶出车道线3031的路口内第一虚拟车道线3041、连接第二驶入车道线3022和第二驶出车道线3032的路口内第二虚拟车道线3042。路口内第一虚拟车道线3041的起点为第一驶入车道线3021与路口停止线305的交点P1、终点为第一驶出车道线3031与路口框301的交点P2，路口内第二虚拟车道线3042的起点为第二驶入车道线3022与路口停止线305的交点P3、终点为第二驶出车道线3032与路口框301的交点P4。路口内第一虚拟车道线3041和路口内第二虚拟车道线3042可以是预先生成的，也可以是实时生成的。依据本申请实施例，可以获得通过该左转弯路口的规划转弯路径。可以理解的，车道线可以是但不限于车道中心线。
参见图2，一种转弯路径规划方法，包括：
在S2001中，根据目标转弯路口的道路信息，获得同向车辆筛选范围。
一实施例中，在自动驾驶的过程中，可以获取自车的定位；根据自车的定位，通过预置的地图例如高精度地图确定自车位于所处的当前道路的具体位置，当检测到自车所处的当前道路包含目标转弯路口(自车在距离该目标转弯路口一定范围时)，可以获取该目标转弯路口的道路信息。目标转弯路口可以是三叉路口，也可以是十字路口或者其他类型的路口。目标转弯路口的道路信息可以包括驶入车道线、驶出车道线、路口的虚拟车道线、停止线、路口范围、驶入车道线的车道边界线、驶出车道线的车道边界线等。在某些实施方式中，自动驾驶车辆也可以通过感知设备(例如摄像头、和/或雷达)获取目标转弯路口的道路信息。
如图3所示，可以根据目标转弯路口的道路信息，获得目标转弯路口的设定范围，即获得目标转弯路口的同向车辆筛选范围。
一个具体实现中，可以对目标转弯路口内最外侧的第二虚拟车道线3042偏移预设距离，获得同向车辆筛选范围的右边界线3061，以连接第一驶入车道321的内侧边界线(第一边界线3211)和第一驶出车道331的内侧边界线(第四边界线3311)的直线为同向车辆筛选范围的左边界线3062。通过同向车辆筛选范围的限制，能够避免目标转弯路口外车辆、障碍物和/或目标转弯路口内的反向行驶车辆对目标车辆筛选的干扰。
在S2002中，将自车前方的、同向车辆筛选范围内符合预设条件的同向车辆确定为目标车辆，并获得目标车辆的多个轨迹点。
一实施例中，在自动驾驶的过程中，自车可以通过感知设备在同向车辆筛选范围采集同向车辆的轨迹信息，同向车辆的轨迹信息包括同向车辆的速度Vagent、航向角θagent、与自车的距离、轨迹点等；通过自车的行驶路径获得自车的航向角θego。根据在同向车辆筛选范围的同向车辆的速度Vagent、航向角θagent、与自车的距离，以及自车的航向角θego，在同向车辆筛选范围对在自车前方的同向车辆进行筛选，将同向车辆与自车的距离不小于设定距离阈值、同向车辆的航向角θagent与自车的航向角θego之差小于设定角度阈值、同向车辆的速度不小于设定速度阈值的同向车辆确定为目标车辆，并获得目标车辆的多个轨迹点。例如，在同向车辆筛选范围内，对在自车前方的、同向车辆进行筛选，将同向车辆的中心与自车的中心距离大于或等于3米的、|θagent-θego|<π/2的、速度Vagent大于或等于2.0米每秒的同向车辆确定为目标车辆，并获得目标车辆的多个轨迹点。
在S2003中，根据目标转弯路口的转弯边界线，建立Frenet坐标系。
一实施例中，目标转弯路口的转弯边界线包括外侧转弯边界线和内侧转弯边界线，外侧转弯边界线为将目标转弯路口的最外侧车道线偏移预设距离获得；目标转弯路口的内侧转弯边界线为连接目标转弯路口的最内侧驶入车道线处的第一边界点和最内侧驶出车道线处的第二边界点的直线。例如，图3所示实施例中，可以以同向车辆筛选范围的右边界线3061为目标转弯路口的外侧转弯边界线，以同向车辆筛选范围的左边界线3062为目标转弯路口的内侧转弯边界线。
一实施例中，以目标转弯路口的外侧转弯边界线为Frenet坐标系的参考线，建立Frenet坐标系，在Frenet坐标系下，沿参考线方向为S轴，垂直于参考线方向为L轴。
在S2004中，在Frenet坐标系下，构建目标转弯路口的外侧转弯边界线和内侧转弯边界线之间的SL网格图。
一实施例中，在Frenet坐标系，采用设定采样步长对目标转弯路口的外侧转弯边界线和内侧转弯边界线之间的同向车辆筛选范围进行采样，构建目标转弯路口的外侧转弯边界线和内侧转弯边界线之间的SL网格图。在Frenet坐标系，设定采样步长包括沿S轴方向的S轴设定采样步长s、沿L轴方向的L轴设定采样步长l。在Frenet坐标系，采用S轴设定采样步长s、L轴设定采样步长l对目标转弯路口的同向车辆筛 选范围进行采样，构建目标转弯路口的设定范围内的SL网格图，SL网格图包括多个网格，各网格的大小为s×l。
一实施例中，采样步长s大于采样步长l。S轴设定采样步长s的范围可以在1.0米～5.0米范围内，例如取值为3.0米。L轴设定采样步长l的范围可以在0.8米～2.0米范围内，例如取值为0.8米。
在S2005中，根据预设路径搜索算法，在SL网格图中获得规划转弯路径的起点和终点之间的多条路径。
一实施例中，目标车辆的多个轨迹点包括转弯路径起点和转弯路径终点。可以根据目标车辆的多个轨迹点，获得规划转弯路径的起点和终点。一些实施例中，如图3所示，可以根据自车驶入目标转弯路口的位置，以点P1或点P3为起点，以点P2或点P4为终点。例如，自车在第一驶入车道321驶入目标转弯路口，以点P1为规划转弯路径的起点、点P2为规划转弯路径的终点。
一实施例中，可以采用Dijkstra(迪杰斯特拉)算法，沿Frenet坐标系的S轴方向在SL网格图中进行搜索，获得转弯路径起点和转弯路径终点之间的多条搜索路径；其中，例如可以将搜索路径所经过的网格的中心点作为搜索路径的轨迹点。
在S2006中，分别获得多条路径中各路径上各网格的设定性能指标函数值。
一实施例中，网格的设定性能指标函数值是根据网格与相邻网格之间的航向角偏移和/或距离偏移获得的。
一实施例中，网格的设定性能指标函数值也称为网格的代价(cost)，包括网格的航向角代价(heading cost)、距离代价(distance cost)，网格的代价用Ggird表示，航向角代价用Gh-c表示，距离代价用Gd-c表示，网格的代价Ggird可至少根据航向角代价Gh-c和距离代价Gd-c获得。多条路径中各路径上各网格的航向角代价，可以根据该网格与下一相邻网格之间的航向角偏移获得；多条路径中各路径上各网格的距离代价，可以根据该网格与下一相邻网格之间的距离偏移获得。
一实施例中，将两个网格中心点的SL坐标转换成笛卡尔坐标系的XY坐标，将两个中心点构成的向量角度作为两个网格之间的航向角偏移。
一实施例中，两个网格中心点之间的欧式距离作为两个网格之间的距离偏移。
在S2007中，根据多个轨迹点在各网格内的分布数据分别确定各网格的权重值。
一实施例中，可以根据多个轨迹点在各网格内的分布数据设置搜索路径中各网格的权重值，例如，设置具有轨迹点的网格的权重高于不具有轨迹点的网格的权重；再例如，设置轨迹点多的网格的权重高于轨迹点少或不具有轨迹点的网格的权重。
一实施例中，根据航向角代价Gh-c和距离代价Gd-c获得网格代价，可以确定多条路径中各网格的代价加权和最小的路径为自车的规划转弯路径。这种情况下，为了使具有轨迹点的网格的权重高于不具有轨迹点的网格的权重，设置具有轨迹点的网格的权重值小于不具有轨迹点的网格的权重值；和/或，设置轨迹点多的网格的权重值小于轨迹点少或不具有轨迹点的网格的权重值。
一实施例中，根据Frenet坐标系的参考线，将目标车辆的多个轨迹点从当前坐标系转换到Frenet坐标系，转换后的多个轨迹点分布于SL网格图的网格内，根据目标车辆的多个轨迹点的Frenet坐标和SL网格中各网格的位置数据，可以获得多个轨迹点在各网格内的分布数据。图4示出SL网格图的一个实例，在SL网格图中，圆点401表示目标车辆的轨迹点，SL网格图中的网格有的不具有目标车辆的轨迹点，有的具有目标车辆的一个轨迹点，有的具有目标车辆的两个轨迹点。
仍以前述根据航向角代价Gh-c和距离代价Gd-c获得网格代价为例，网格的权重值也可称为网格的衰减因子Fagent，衰减因子Fagent是小于1或等于1的系数。具有目标车辆的轨迹点的网格的衰减因子Fagent小于1，不具有目标车辆的轨迹点的网格的衰减 因子Fagent为1.0。在一个具体实例中，具有目标车辆的至少两个轨迹点的网格的衰减因子Fagent为0.6，具有目标车辆的一个轨迹点的网格的衰减因子Fagent为0.8，不具有目标车辆的轨迹点的网格的衰减因子Fagent为1.0。通过设置具有目标车辆的轨迹点的网格的衰减因子Fagent小于1，可以减少具有目标车辆的轨迹点的各网格的代价。
一实施例中，对于SL网格图的第i个网格，其加权代价Ggird-i＝Fagent-i*(Gh-c-i+Gd-c-i)，Fagent-i、Gh-c-i、Gd-c-i分别为SL网格图中第i个网格的衰减因子、航向角代价、距离代价。
在S2008中，根据目标转弯路口的形状数据及各网格的位置数据，获得各网格权重值的修正系数。
一实施例中，为使自车根据规划转弯路径在目标转弯路口移动时尽量避免目标转弯路口的车道边界，可以根据目标转弯路口的长度K、宽度W的比例关系，以及SL网格图各网格的L轴数据，分别获得各网格相对于设定安全位置的偏移距离；根据各网格的偏移距离获得各网格权重值的修正系数。
如图4所示，沿Frenet坐标系的L轴403方向上设计修正系数曲线402，修正系数可以使更靠近设定安全位置的网格具有更高的权重。
一实施例中，SL网格图第i个网格的修正系数为FL-i，对于SL网格图的第i个网格，其修正后的加权代价Ggird-i＝FL-i*Fagent-i*(Gh-c-i+Gd-c-i)。
确定多条路径中各网格的代价加权和最小的路径为自车的规划转弯路径的实施例中，修正系数FL-i可以是大于或等于1的系数，且网格越靠近设定安全中心，修正系数越接近1。
一实施例中，修正系数曲线402为沿Frenet坐标系的L轴方向的近抛物线，其中心对应于SL网格图的L轴中心位置，可表示为Lx/2，该中心位置对应的极值为1.0。由于不同路目标转弯路口的形状有所不同，可利用路口的长宽比例调整修正系数曲线的中心位置。
一实施例中，可以通过以下公式计算SL网格图第i个网格的权重值的修正系数FL-i：
其中，Li为SL网格图第i个网格的L轴数据，可以通过i乘以L轴采样步长l获得；为第i个网格相对于设定安全位置的偏移距离；Lx/2为SL网格图的L轴的中心位置。
在S2009中，根据各网格的性能指标函数值、权重值和修正系数，获得多条路径中的每条路径的评价函数值。
一实施例中，在Frenet坐标系，多条路径中的各路径经过SL网格图中的多个网格，连接规划转弯路径的起点和终点。将各路径上各网格的设定性能指标函数值加权相加，以各路径上各网格的设定性能指标函数值的加权和为多条路径中的每条路径的评价函数值。
一实施例中，根据各网格的权重值和修正系数，各网格修正后的加权代价Ggird-i＝FL-i*Fagent-i*(Gh-c-i+Gd-c-i)。多条路径中的每条路径的评价函数值可以是多条路径中的每条路径的多个网格的修正后的加权代价的和∑Ggird。例如，多条路径中的一条路径经过N个网格，N个网格中各网格的修正后的加权代价根据Ggird-i＝FL-i*Fagent-i*(Gh-c-i+Gd-c-i)获得；将N个网格中各网格的修正后的加权代价相加获得N个网格的加权代价和∑Ggird，该N个网格的加权代价和∑Ggird即为该条路径的评价函数值。
在S2010中，根据多条路径各自的评价函数值，确定多条路径中评价函数值符合 预设条件的路径为自车的规划转弯路径，以使自车根据规划转弯路径移动通过目标转弯路口。
一实施例中，可以根据Frenet坐标系多条路径中各路径的评价函数值，选择评价函数值最小的路径作为自车通过目标转弯路口的Frenet坐标系的规划转弯路径，以使自车根据规划转弯路径移动通过目标转弯路口。
一实施例中，可以将Frenet坐标系的规划转弯路径，通过坐标系转换，获得笛卡尔坐标系的规划转弯路径，以使自车根据笛卡尔坐标系的规划转弯路径移动通过目标转弯路口。
一实施例中，可以通过平滑算法对自车的规划转弯路径进行平滑处理，获得通过目标转弯路口的平滑的规划转弯路径。一些实施例中，可以通过包括但不限于CG(Conjugate Gradient，共轭梯度)算法，对自车的规划转弯路径进行平滑处理，获得平滑的规划转弯路径，以使自车根据平滑的规划转弯路径平滑移动通过目标转弯路口。采用CG算法的平滑处理，实现简单，计算耗时小。
本申请实施例中，根据自车前方的目标车辆通过目标转弯路口的转弯路径的多个轨迹点，获得自车的规划转弯路径，以使自车根据规划转弯路径移动通过目标转弯路口；自车可以不必识别目标转弯路口的车道线，根据目标转弯路口的实际车流，根据自车前方的目标车辆通过目标转弯路口的转弯路径的多个轨迹点进行转弯路径规划，获得自车的规划转弯路径，目标转弯路口内目标车辆行驶的转弯路径可以避开目标转弯路口中的危险地段或障碍物，自车在根据规划转弯路径通过目标转弯路口时，也能避开目标转弯路口中的危险地段或障碍物，能够避免自车的危险切入或者碰撞，提高车辆通过目标转弯路口的安全性。
进一步的，本申请实施例中，将目标转弯路口内的转弯边界线作为Frenet坐标系的参考线，构建Frenet坐标系；在Frenet坐标系构建目标转弯路口的设定范围内的SL网格图，将目标车辆的多个轨迹点从当前坐标系转换到Frenet坐标系，多个轨迹点分布于SL网格图的网格内，根据多个轨迹点在SL网格图各网格内的分布数据确定各网格的权重值，设置具有轨迹点的网格的权重值小于不具有轨迹点的网格的权重值；和/或，设置轨迹点多的网格的权重值小于轨迹点少或不具有轨迹点的网格的权重值，通过各网格权重值的设置，减少具有目标车辆的轨迹点的各网格的代价，使得自车的规划转弯路径尽可能的经过目标车辆的轨迹点，能够避免自车的危险切入或者碰撞，提高车辆通过目标转弯路口的安全性。
进一步的，本申请实施例中，根据目标转弯路口的长度和宽度的比例关系、及各网格的L轴数据，分别获得各网格相对于设定安全位置的偏移数据；根据各网格的偏移数据获得各网格的修正系数，修正系数是大于1的系数，越靠近目标转弯路口中心位置的各网格的修正系数越小，使得越靠近中心位置的各网格的代价越小，则自车的规划转弯路径越靠近目标转弯路口的设定范围的中心位置，能够避免自车的规划转弯路径位于目标转弯路口的边界，自车根据规划转弯路径移动通过目标转弯路口时，能够偏离目标转弯路口两侧的边界，能够避免自车的危险切入或者碰撞，提高车辆通过目标转弯路口的安全性。
与前述应用功能实现方法实施例相对应，本申请还提供了一种计算设备及相应的实施例。
图5是本申请一实施例示出的计算设备的结构示意图。
计算设备可以是一个或多个计算机终端或可以是服务器，或者也可以是计算机终端和服务器的组合等。可以理解的，服务器可以是一个物理服务器或者多个物理服务器虚拟而成的一个逻辑服务器。服务器也可以是多个可互联通信的服务器组成的服务器群，且各个功能模块可分别分布在服务器群中的各个服务器上。一些实施例中，计 算设备是车载电子设备，例如可以是但不限于车辆的电子控制单元、自动驾驶系统控制器、智能导航设备、智能手机、智能平板设备等可移动设备等。
参见图5，计算设备500包括存储器501和处理器502。
可以理解的，本申请中，计算设备500包括处理器502及存储有计算机程序的存储器501。上述处理器502执行所存储的计算机程序时可以实现转弯路径规划方法。计算设备500可以是一个或多个计算机终端或可以是服务器，或者也可以是计算机终端和服务器的组合等。可以理解的，服务器可以是一个物理服务器或者多个物理服务器虚拟而成的一个逻辑服务器。服务器也可以是多个可互联通信的服务器组成的服务器群，且各个功能模块可分别分布在服务器群中的各个服务器上。一些实施例中，计算设备500是车载电子设备，例如可以是但不限于车辆的电子控制单元、自动驾驶系统控制器、智能导航设备、智能手机、智能平板设备等可移动设备等。
处理器502可以是中央处理单元(Central Processing Unit，CPU)，还可以是其他通用处理器、数字信号处理器(Digital Signal Processor，DSP)、专用集成电路(Application Specific Integrated Circuit，ASIC)、现场可编程门阵列(Field-Programmable Gate Array，FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件等。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。
存储器501可以包括各种类型的存储单元，例如系统内存、只读存储器(ROM)和永久存储装置。其中，ROM可以存储处理器502或者计算机的其他模块需要的静态数据或者指令。永久存储装置可以是可读写的存储装置。永久存储装置可以是即使计算机断电后也不会失去存储的指令和数据的非易失性存储设备。在一些实施方式中，永久性存储装置采用大容量存储装置(例如磁或光盘、闪存)作为永久存储装置。另外一些实施方式中，永久性存储装置可以是可移除的存储设备(例如软盘、光驱)。系统内存可以是可读写存储设备或者易失性可读写存储设备，例如动态随机访问内存。系统内存可以存储一些或者所有处理器在运行时需要的指令和数据。此外，存储器501可以包括任意计算机可读存储媒介的组合，包括各种类型的半导体存储芯片(例如DRAM，SRAM，SDRAM，闪存，可编程只读存储器)，磁盘和/或光盘也可以采用。在一些实施方式中，存储器501可以包括可读和/或写的可移除的存储设备，例如激光唱片(CD)、只读数字多功能光盘(例如DVD-ROM，双层DVD-ROM)、只读蓝光光盘、超密度光盘、闪存卡(例如SD卡、min SD卡、Micro-SD卡等)、磁性软盘等。计算机可读存储媒介不包含载波和通过无线或有线传输的瞬间电子信号。
存储器501上存储有可执行代码，当可执行代码被处理器502处理时，可以使处理器502执行上文述及的方法中的部分或全部。
本申请还提供一种车辆，包括如上所述的计算设备500。
此外，根据本申请的方法还可以实现为一种计算机程序或计算机程序产品，该计算机程序或计算机程序产品包括用于执行本申请的上述方法中部分或全部步骤的计算机程序代码指令。
或者，本申请还可以实施为一种计算机可读存储介质(或非暂时性机器可读存储介质或机器可读存储介质)，其上存储有可执行代码(或计算机程序或计算机指令代码)，当可执行代码(或计算机程序或计算机指令代码)被车辆(或计算设备、服务器等)的处理器执行时，使处理器执行根据本申请的上述方法的各个步骤的部分或全部。
以上已经描述了本申请的各实施例，上述说明是示例性的，并非穷尽性的，并且也不限于所披露的各实施例。在不偏离所说明的各实施例的范围和精神的情况下，对于本技术领域的普通技术人员来说许多修改和变更都是显而易见的。本文中所用术语 的选择，旨在最好地解释各实施例的原理、实际应用或对市场中的技术的改进，或者使本技术领域的其他普通技术人员能理解本文披露的各实施例。

Claims (13)

  1. 一种转弯路径规划方法，其中，所述方法包括：
    获得自车前方的目标车辆在目标转弯路口内的转弯路径的多个轨迹点，所述多个轨迹点包括转弯路径起点和转弯路径终点；
    获得所述转弯路径起点和转弯路径终点之间的多条路径；
    根据所述目标车辆的所述多个轨迹点，确定所述多条路径中评价函数值符合预设条件的路径为所述自车的规划转弯路径，以使所述自车根据所述规划转弯路径移动通过所述目标转弯路口。
  2. 根据权利要求1所述的方法，其中，
    所述根据所述目标车辆的所述多个轨迹点，确定所述多条路径中评价函数值符合预设条件的路径为所述自车的规划转弯路径，包括：
    确定Frenet坐标系的参考线，其中，所述参考线为所述目标转弯路口内的转弯边界线；
    根据所述参考线，将所述多个轨迹点从当前坐标系转换到Frenet坐标系；
    在所述Frenet坐标系下，根据所述多个轨迹点和预设的评价函数，分别获得所述转弯路径起点和转弯路径终点之间的多条路径各自的评价函数值；
    根据所述多条路径各自的评价函数值，确定所述多条路径中评价函数值符合预设条件的路径为所述自车的规划转弯路径。
  3. 根据权利要求2所述的方法，其中，
    所述获得所述转弯路径起点和转弯路径终点之间的多条路径，包括：构建所述目标转弯路口的设定范围内的SL网格图，所述多个轨迹点分布于所述SL网格图的网格内；根据预设路径搜索算法，在所述SL网格图中获得所述转弯路径起点和转弯路径终点之间的多条路径；
    所述获得所述多条路径中的每条路径的评价函数值，包括：
    分别获得所述路径上各网格的设定性能指标函数值；
    根据所述多个轨迹点在所述各网格内的分布数据分别确定所述各网格的权重值；
    根据所述各网格的所述性能指标函数值和所述权重值，获得所述多条路径中的每条路径的评价函数值。
  4. 根据权利要求3所述的方法，其中，
    所述参考线为所述目标转弯路口的外侧转弯边界线，所述外侧转弯边界线为将所述目标转弯路口的最外侧车道线偏移预设距离获得；
    所述目标转弯路口的内侧转弯边界线为连接所述目标转弯路口的最内侧驶入车道线处的第一边界点和最内侧驶出车道线处的第二边界点的直线；
    所述构建所述目标转弯路口的设定范围内的SL网格图，包括：构建所述目标转弯路口的外侧转弯边界线和内侧转弯边界线之间的SL网格图。
  5. 根据权利要求3所述的方法，其中，所述网格的设定性能指标函数值是根据所述网格与相邻网格之间的航向角偏移和/或距离偏移获得的。
  6. 根据权利要求3所述的方法，其中，所述根据所述多个轨迹点在所述各网格内的分布数据确定所述各网格的权重值，包括：
    设置具有轨迹点的网格的权重高于不具有轨迹点的网格的权重；和/或，
    设置轨迹点多的网格的权重高于轨迹点少或不具有轨迹点的网格的权重。
  7. 根据权利要求3所述的方法，其中，所述根据所述各网格的所述性能指标函数值和所述权重值，获得所述多条路径中的每条路径的评价函数值，包括：
    根据所述目标转弯路口的形状数据及各网格的位置数据，获得所述各网格的权重值的修正系数；
    根据所述各网格的所述性能指标函数值、所述权重值和所述修正系数，获得所述多条路径中的每条路径的评价函数值。
  8. 根据权利要求7所述的方法，其中，所述根据所述目标转弯路口的形状数据及所述各网格的位置数据，获得所述各网格的权重值的修正系数，包括：
    根据所述目标转弯路口的长度和宽度的比例关系、及所述各网格的L轴数据，分别获得所述各网格相对于设定安全位置的偏移数据；
    根据所述各网格的所述偏移数据获得所述各网格的权重值的修正系数。
  9. 根据权利要求1至8中任一项所述的方法，其中，所述获得自车前方的目标车辆在目标转弯路口内的转弯路径的多个轨迹点，还包括：
    将自车前方处于设定路口范围内、且符合预设条件的同向车辆确定为目标车辆，所述符合预设条件包括以下部分或全部：同向车辆与自车之间的距离不小于设定距离阈值；同向车辆的航向角与自车的航向角之差小于设定角度阈值；同向车辆的速度不小于设定速度阈值。
  10. 根据权利要求9所述的方法，其中，所述方法还包括：通过共轭梯度算法对所述自车的规划转弯路径进行平滑处理。
  11. 一种计算设备，其中，所述计算设备包括：
    处理器；以及
    存储器，其上存储有可执行代码，当所述可执行代码被所述处理器执行时，使所述处理器执行如权利要求1至10中任一项所述的方法。
  12. 一种车辆，其中，具有如权利要求11所述的计算设备。
  13. 一种计算机可读存储介质，其中，其上存储有可执行代码，当所述可执行代码被处理器执行时，使所述处理器执行如权利要求1至10中任一项所述的方法。
PCT/CN2023/119059 2022-09-16 2023-09-15 转弯路径规划方法、设备、车辆及存储介质 WO2024056064A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211125386.XA CN115200604B (zh) 2022-09-16 2022-09-16 转弯路径规划方法、设备、车辆及存储介质
CN202211125386.X 2022-09-16

Publications (1)

Publication Number Publication Date
WO2024056064A1 true WO2024056064A1 (zh) 2024-03-21

Family

ID=83571746

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/119059 WO2024056064A1 (zh) 2022-09-16 2023-09-15 转弯路径规划方法、设备、车辆及存储介质

Country Status (2)

Country Link
CN (1) CN115200604B (zh)
WO (1) WO2024056064A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115200604B (zh) * 2022-09-16 2023-05-16 广州小鹏自动驾驶科技有限公司 转弯路径规划方法、设备、车辆及存储介质
CN117109619A (zh) * 2023-08-28 2023-11-24 北京工业大学 一种无人驾驶路径规划方法
CN117451072B (zh) * 2023-12-21 2024-03-26 广汽埃安新能源汽车股份有限公司 一种车辆起步阶段的轨迹规划方法及装置

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108995708A (zh) * 2017-06-06 2018-12-14 蔚来汽车有限公司 一种车辆自动转弯控制方法和装置
CN112068545A (zh) * 2020-07-23 2020-12-11 哈尔滨工业大学（深圳） 一种无人驾驶车辆在十字路口的行驶轨迹规划方法、系统及存储介质
CN112683291A (zh) * 2020-12-30 2021-04-20 广州小鹏自动驾驶科技有限公司 车辆转弯路径规划方法、装置、车辆和存储介质
US20210221374A1 (en) * 2018-09-28 2021-07-22 Baidu Usa Llc Spiral path based three-point turn planning for autonomous driving vehicles
CN113291318A (zh) * 2021-05-28 2021-08-24 同济大学 基于部分可观测马尔科夫模型的无人车盲区转弯规划方法
CN114834486A (zh) * 2022-06-08 2022-08-02 云控智行科技有限公司 一种规划车辆掉头路径的方法、装置及设备
CN115200604A (zh) * 2022-09-16 2022-10-18 广州小鹏自动驾驶科技有限公司 转弯路径规划方法、设备、车辆及存储介质

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9696720B2 (en) * 2015-08-27 2017-07-04 Hyundai Motor Company Apparatus and method for controlling autonomous navigation
US11506505B2 (en) * 2019-02-13 2022-11-22 The Boeing Company Methods and apparatus for determining a vehicle path
CN112660148B (zh) * 2019-09-30 2022-09-02 阿波罗智能技术(北京)有限公司 确定车辆的调头路径的方法、装置、设备和介质
CN111857160B (zh) * 2020-08-17 2023-06-02 武汉中海庭数据技术有限公司 一种无人车路径规划方法和装置
CN113932823A (zh) * 2021-09-23 2022-01-14 同济大学 基于语义道路地图的无人驾驶多目标点轨迹并行规划方法
CN114485695A (zh) * 2021-12-20 2022-05-13 北京罗克维尔斯科技有限公司 路径规划方法、装置、服务器、车辆及存储介质
CN114763996A (zh) * 2022-04-27 2022-07-19 重庆长安汽车股份有限公司 一种基于多传感器融合的复杂场景路径规划方法
CN114964293A (zh) * 2022-05-31 2022-08-30 智道网联科技(北京)有限公司 车辆的路径规划方法、装置及电子设备、存储介质

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108995708A (zh) * 2017-06-06 2018-12-14 蔚来汽车有限公司 一种车辆自动转弯控制方法和装置
US20210221374A1 (en) * 2018-09-28 2021-07-22 Baidu Usa Llc Spiral path based three-point turn planning for autonomous driving vehicles
CN112068545A (zh) * 2020-07-23 2020-12-11 哈尔滨工业大学（深圳） 一种无人驾驶车辆在十字路口的行驶轨迹规划方法、系统及存储介质
CN112683291A (zh) * 2020-12-30 2021-04-20 广州小鹏自动驾驶科技有限公司 车辆转弯路径规划方法、装置、车辆和存储介质
CN113291318A (zh) * 2021-05-28 2021-08-24 同济大学 基于部分可观测马尔科夫模型的无人车盲区转弯规划方法
CN114834486A (zh) * 2022-06-08 2022-08-02 云控智行科技有限公司 一种规划车辆掉头路径的方法、装置及设备
CN115200604A (zh) * 2022-09-16 2022-10-18 广州小鹏自动驾驶科技有限公司 转弯路径规划方法、设备、车辆及存储介质

Also Published As

Publication number Publication date
CN115200604B (zh) 2023-05-16
CN115200604A (zh) 2022-10-18

Similar Documents

Publication Publication Date Title
WO2024056064A1 (zh) 转弯路径规划方法、设备、车辆及存储介质
AU2017418043B2 (en) Systems and methods for trajectory determination
CN110155053B (zh) 提供用于驾驶车辆的信息的方法和设备
JP6644742B2 (ja) 頑健で効率的な車両測位用のアルゴリズム及びインフラ
EP3109842B1 (en) Map-centric map matching method and apparatus
US10579062B2 (en) Scalable smooth reference path generator for autonomous driving vehicles
US10824153B2 (en) Cost design for path selection in autonomous driving technology
JP2020132151A (ja) 車両走行軌跡の予測方法、装置、端末、記憶媒体、及びプログラム
KR102106359B1 (ko) 전자 지도 중의 교차로를 인식하기 위한 방법 및 장치
WO2020108207A1 (zh) 一种高精度地图的构建方法及装置
CN109489673A (zh) 用于自动驾驶车辆的数据驱动地图更新系统
CN110531749A (zh) 确定避开移动障碍物的自动驾驶的驾驶路径
CN106541945A (zh) 一种基于icp算法的无人车自动泊车方法
CN108995657A (zh) 操作自动驾驶车辆的方法和数据处理系统
CN108106620B (zh) 一种拓扑道路匹配方法、系统及电子设备
CN112444263B (zh) 全局路径规划方法及装置
JP2015082326A (ja) 交差点における画像及び地図に基づく車両の検出法
CN108437972B (zh) 一种基于位置偏差的轨迹跟踪方法和装置
JP7234354B2 (ja) 走行座標系の構築方法及びその使用
CN108961811A (zh) 停车场车辆定位方法、系统、移动终端及存储介质
CN108332761B (zh) 一种使用及创建路网地图信息的方法与设备
JP2020132155A (ja) 運転参照経路の処理方法、装置、車両、及びプログラム
CN109491378A (zh) 自动驾驶车辆的基于道路分段的路线引导系统
KR102534412B1 (ko) 환경 목표의 특징점을 추출하는 방법 및 장치
CN109683617A (zh) 一种自动驾驶方法、装置及电子设备

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23864793

Country of ref document: EP

Kind code of ref document: A1